European Journal of Nuclear Medicine and Molecular Imaging

, Volume 33, Issue 5, pp 557–565

In vitro validation of bioluminescent monitoring of disease progression and therapeutic response in leukaemia model animals

Authors

    • Department of Radiology, Institute of Medical ScienceUniversity of Tokyo
  • Arinobu Tojo
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Rieko Sekine
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Yasushi Soda
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Seiichiro Kobayashi
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Akiko Nomura
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Kiyoko Izawa
    • Division of Molecular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Toshio Kitamura
    • Division of Cellular Therapy, Advanced Clinical Research CentreUniversity of Tokyo
  • Toshiyuki Okubo
    • Department of Radiology, Institute of Medical ScienceUniversity of Tokyo
  • Kuni Ohtomo
    • Department of Radiology, Graduate School of MedicineUniversity of Tokyo
Molecular imaging

DOI: 10.1007/s00259-005-0048-4

Cite this article as:
Inoue, Y., Tojo, A., Sekine, R. et al. Eur J Nucl Med Mol Imaging (2006) 33: 557. doi:10.1007/s00259-005-0048-4

Abstract

Purpose

The application of in vivo bioluminescence imaging to non-invasive, quantitative monitoring of tumour models relies on a positive correlation between the intensity of bioluminescence and the tumour burden. We conducted cell culture studies to investigate the relationship between bioluminescent signal intensity and viable cell numbers in murine leukaemia model cells.

Methods

Interleukin-3 (IL-3)-dependent murine pro-B cell line Ba/F3 was transduced with firefly luciferase to generate cells expressing luciferase stably under the control of a retroviral long terminal repeat. The luciferase-expressing cells were transduced with p190 BCR-ABL to give factor-independent proliferation. The cells were cultured under various conditions, and bioluminescent signal intensity was compared with viable cell numbers and the cell cycle stage.

Results

The Ba/F3 cells showed autonomous growth as well as stable luciferase expression following transduction with both luciferase and p190 BCR-ABL, and in vivo bioluminescence imaging permitted external detection of these cells implanted into mice. The bioluminescence intensities tended to reflect cell proliferation and responses to imatinib in cell culture studies. However, the luminescence per viable cell was influenced by the IL-3 concentration in factor-dependent cells and by the stage of proliferation and imatinib concentration in factor-independent cells, thereby impairing the proportionality between viable cell number and bioluminescent signal intensity. Luminescence per cell tended to vary in association with the fraction of proliferating cells.

Conclusion

Although in vivo bioluminescence imaging would allow non-invasive monitoring of leukaemia model animals, environmental factors and therapeutic interventions may cause some discrepancies between tumour burden and bioluminescence intensity.

Keywords

LuciferaseLeukaemiaRetrovirusesImatinib mesylateCell cycle

Introduction

In vivo bioluminescence imaging is used increasingly to evaluate the effects of novel therapeutic strategies against malignant neoplasms in small animal models [1, 2]. For monitoring by bioluminescence imaging, mice are inoculated with tumour cells that stably express luciferase under the control of a constitutive promoter, such as the simian virus 40 (SV40) promoter, the cytomegalovirus (CMV) immediate-early promoter or the long-terminal repeat (LTR) of a retrovirus. Injection of the mice with luciferin, substrate for luciferase, induces light emission from the luciferase-expressing cells, and images that reflect the amount and whole-body distribution of the implanted cells can be acquired using a charge-coupled device (CCD) camera. Quantitative indices of tumour burden can be computed from the images. Owing to the convenient, non-invasive nature of the imaging procedures, measurements can be performed repetitively to assess tumour growth and therapeutic efficacy using each animal as its own control.

The monitoring of tumour models by in vivo bioluminescence imaging relies on a positive correlation between signal intensity on bioluminescence imaging and tumour burden. Ideally, changes in viable cell number result in proportional changes in light emission. In vitro experiments have demonstrated a linear relationship between cell numbers and light emission after the addition of D-luciferin to a dilution series of luciferase-expressing cells [37]. It has also been reported that the signal intensities obtained by in vivo bioluminescence imaging correlate positively with tumour burden in various animal models [3, 6, 819]. These observations support the use of bioluminescence imaging for quantitative evaluations of implanted tumour progression and regression. On the other hand, the activity of the CMV promoter has been shown to depend on the cell cycle stage and medium composition [20], and expression driven by the LTR has been suggested to decline under a stress condition [21]. It may be possible that alterations in the physiological status of the cell cause fluctuations in luciferase expression under the control of a constitutive promoter, thereby distorting the proportionality between viable cell number and bioluminescent signal intensity. It has not been fully examined whether luciferase activity increases with increasing viable cell number during disease progression or whether therapeutic interventions affect the level of luciferase activity per viable cell.

The Philadelphia chromosome (Ph) contains one of several types of BCR-ABL fusion gene and is important in the pathogenesis of both acute lymphoblastic leukaemia (ALL) and chronic myeloid leukaemia [22]. The BCR-ABL fusion proteins retain constitutive tyrosine kinase activity, leading to uncontrolled cell proliferation. Patients with Ph+ ALL frequently express the p190 BCR-ABL fusion protein and have a very poor prognosis [2325]. Although the BCR-ABL tyrosine kinase inhibitor, imatinib mesylate (STI571; Novartis Pharmaceuticals, Basel, Switzerland), is effective in the treatment of Ph+ ALL patients [26], resistance to this drug develops rapidly, and novel therapeutic strategies to overcome the resistance need to be explored [27]. Since tumour cells may be distributed extensively and variably in leukaemia model animals, making it difficult to assess disease severity, whole-body, quantitative evaluation of tumour burden by in vivo bioluminescence imaging appears to have particular value [28].

The interleukin-3 (IL-3)-dependent murine pro-B cell line Ba/F3 [29] shows autonomous proliferation following transduction with the p190 BCR-ABL fusion gene [30]. In this study, we generated p190 BCR-ABL-transformed Ba/F3 cells stably expressing luciferase under the control of a retroviral LTR. Cell culture studies were conducted to investigate whether bioluminescent signal intensities could be used as indicators of cell proliferation and responses to imatinib. We cultured factor-dependent cells in the presence of different concentrations of IL-3 and measured the luciferase activities and viable cell numbers to evaluate the effect of IL-3 on luciferase expression. The proliferation of factor-independent cells was monitored serially by the luciferase assay and viable cell counting, and the relationship between them was defined, relative to the stage of proliferation. The effects of imatinib on luciferase activity and viable cell number were assessed to evaluate the reliability of bioluminescent monitoring of therapeutic responses. Our results indicate that changes in bioluminescent signal intensity generally reflect cell proliferation and therapeutic responses but differ, to some extent, from changes in viable cell number depending on cell conditions associated with proliferative activity.

Materials and methods

Cell lines

Ba/F3 cells were maintained in RPMI 1640 medium (Invitrogen, Grand Island, NY, USA) supplemented with 10% (v/v) fetal bovine serum (FBS; JRH Biosciences, Lenexa, KS, USA), 1% penicillin/streptomycin (Invitrogen) and 100–200 pg/ml recombinant murine IL-3 (mIL-3; kindly provided by Kirin Brewery, Maebashi, Japan). Ba/F3 cells transduced with the BCR-ABL genes were cultured in the absence of mIL-3. The culture density was kept below 5×105 cells/ml. A retrovirus-packaging cell line for ecotropic retroviruses, Plat-E [31], was maintained in Dulbecco’s modified Eagle’s medium (Invitrogen) supplemented with 10% FBS and 1% penicillin/streptomycin. The medium also contained 1 μg/ml puromycin (Sigma Chemical Co., St Louis, MO, USA) and 10 μg/ml blasticidin S (Funakoshi Co., Tokyo, Japan) as selection reagents. All of the cultures were incubated at 37°C and 5% CO2.

Construction of plasmids and retroviral transduction

The cDNA encoding the firefly luciferase was excised from the pGL3-basic vector (Promega, Madison, WI, USA) and inserted into the retroviral vector pMX-neo [32], to generate pMX-luc/neo. The pMX-neo employs the LTR of Moloney murine leukaemia virus (MMLV) for the expression of inserted sequence and harbours a SV40 early promoter-driven neomycin resistance gene. The wild-type and mutant p190 BCR-ABL fusion genes were inserted into the retroviral vector pMC-Ig [32], to generate pMC-p190wt/Ig and pMC-p190mut/Ig, respectively. The pMC-Ig contains the enhanced green fluorescence protein (EGFP) gene downstream of the internal ribosome entry site. The mutant gene, which harbours the Y253H point mutation in the BCR-ABL kinase domain, was constructed by replacing the kinase domain of the wild-type cDNA with the corresponding mutated sequence derived from leukaemia cells from an imatinib-resistant Ph+ ALL patient [33].

The luciferase expression plasmid pMX-luc/neo was transfected into Plat-E cells to generate the ecotropic retroviral vector. Plat-E cells (1.5×106 cells/3 ml) were seeded in a 60-mm dish, and pMX-luc/neo was transfected 16 h later using FuGENE 6 Transfection Reagent (Roche Diagnostics, Indianapolis, IN, USA) according to the manufacturer’s protocol. The culture supernatants were harvested 48 and 72 h after transfection, and Ba/F3 cells were transduced in the presence of polybrene. The infected Ba/F3 cells were selected for 14 days with 1.0 mg/ml G418 (Calbiochem, San Diego, CA, USA) and termed Ba/F3-Luc cells. Similarly, the Ba/F3-Luc cells and parenteral Ba/F3 cells were transduced with pMC-p190wt/Ig or pMC-p190mut/Ig. The infected cells were selected by IL-3 depletion for 14 days. The Ba/F3-Luc cells transduced with the wild-type and mutant p190 genes were referred to as Ba/F3-Luc/Wt and Ba/F3-Luc/Mut cells, respectively.

In vitro analysis

The standard luciferase assay, intact-cell luciferase assay, viable cell counting and cell cycle analysis were performed to assess the bioluminescent features and proliferative status of cultured cells. All measurements were done in triplicate. The viable cell numbers were measured using the trypan blue dye exclusion method and a haemocytometer. For comparisons of sensitivities to imatinib between cell lines, a cell titre assay was performed using the WST-8 assay kit (TetraColor One; Seikagaku Co., Tokyo, Japan) according to the manufacturer’s recommendations.

Luciferase activity in a given volume of cell suspension was determined by the standard luciferase assay. To prepare the lysate, 100 μl of cell suspension was transferred from a cell culture plate to a microtube and centrifuged on a tabletop centrifuge (2,000 rpm, 5 min). The pellet was lysed with 200 μl of lysis buffer (Passive Lysis Buffer; Promega). The lysate was centrifuged, and the supernatant was stored at −80°C until assayed. Luminescence from the lysate was measured using the Luciferase Assay Reagent (Promega) according to the manufacturer’s recommendation and using a plate reader (Wallac ARVO MX 1420 Multilabel Counter; Perkin Elmer Japan, Yokohama, Japan). In some experiments, luminescence was also measured by simply adding D-luciferin (Beetle Luciferin Potassium Salt; Promega) to the cell suspension without cell lysis. We referred to this latter assay as the intact-cell luciferase assay. Cell suspension (50 μl) was transferred to a white 96-well cell culture plate. One minute after the addition of D-luciferin (10 μl of 600 μg/ml solution) to the cell suspension, the light output was measured using the plate reader. Phenol red-free RPMI 1640 medium was used to avoid possible light absorbance by the dye. For both the standard and the intact-cell luciferase assay, luminescence per cell (counts per second/cell; cps/cell) was calculated from the mean luminescence and mean viable cell number. The measurements of luminescence using the plate reader were performed at 25°C.

To assess the cell cycle, cells were fixed with cooled 70% ethanol. Afterwards, the fixed cells were washed twice with phosphate-buffered saline and incubated with 0.5% ribonuclease A for 30 min. After the addition of propidium iodide (final concentration, 50 μg/ml), the cells were analysed by flow cytometry using the FACSCalibur flow cytometer (Beckton Dickinson, Franklin Lakes, NJ, USA). The cell cycle was analysed using the FlowJo software (TreeStar, San Carlos, CA, USA). The fraction of proliferating cells, or proliferation index, was calculated by the following equation: proliferation index (%)=(G2/M+S)/(G1/G0+G2/M+S)×100, where G2/M, S and G1/G0 are the numbers of cells in the G2/M, S and G1/G0 phases, respectively.

In vivo bioluminescence imaging

Two female wild-type BALB/c mice were inoculated subcutaneously in the right femoral region with 1×105 Ba/F3-Luc/Wt cells. Five minutes later, the mice received an intraperitoneal injection of 150 mg/kg D-luciferin and placed in the light-tight chamber of a cooled CCD camera system (IVIS Imaging System 100; Xenogen, Alameda, CA, USA) in the prone position under isoflurane anaesthesia. Photographic and luminescent images were acquired 20 min after D-luciferin injection using the CCD camera system. In addition, two female BALB/c nu/nu mice were injected intravenously with 2×106 Ba/F3-Luc/Wt cells, followed 10 min later with injection of D-luciferin. Dorsal, left lateral, ventral and right lateral images were acquired from 10 min after D-luciferin injection with the CCD camera system. All luminescent images were collected with an exposure time of 1 min and binning of 8. Mice were handled according to the guidelines of the Institute of Medical Science, University of Tokyo. The experiments were approved by the committee for animal research at the institution.

Results

Generation of factor-independent Ba/F3 cell lines expressing luciferase

We transduced Ba/F3 cells with firefly luciferase genes using a retroviral vector, then selected with G418, and confirmed that the cells expressed luciferase. We then transduced the obtained luciferase-expressing Ba/F3-Luc cells and parental Ba/F3 cells with wild-type or mutant p190 BCR-ABL fusion genes, to give IL-3-independent, autonomous cellular proliferation. A clone from each of the four cell lines was selected for further investigation based on the presence of a single peak of EGFP expression on flow cytometry and, for luciferase-expressing cells, strong expression of luciferase.

The cell growth curves, determined by viable cell counting, were similar for the four cell lines (data not shown), and no significant effect of luciferase expression on the proliferation rate was noted. For the evaluation of sensitivity to imatinib, we cultured the four cell lines in the presence of various concentrations of imatinib (0–10 μM) and performed the standard cell titre assay. The dose-response curves demonstrated the sensitivity of Ba/F3-Luc/Wt cells and resistance of Ba/F3-Luc/Mut cells to imatinib (data not shown). The transduction with luciferase genes did not influence sensitivity to imatinib. Furthermore, luciferase expression by Ba/F3-Luc/Wt cells maintained in the absence of G418 was assessed repeatedly. The standard luciferase assay and viable cell counting were performed 24 h after replating. The luminescence per cell remained constant (range 159.6–167.0 cps/cell) from 5 to 39 days after thawing the frozen cell stock, which indicates excellent long-term stability of luciferase expression even in the absence of selection pressure.

In vivo bioluminescence imaging

We examined the detectability of the luciferase-expressing cells by in vivo bioluminescence imaging. For the mice inoculated subcutaneously with 1×105 Ba/F3-Luc/Wt cells, light emission was clearly detected at the site of cell implantation (Fig. 1a). For the mice injected intravenously with 2×106 Ba/F3-Luc/Wt cells, light emission was detected throughout the body (Fig. 1b,c), indicating diffuse distribution of the injected cells. Relatively strong signals were shown for the lung, liver and spleen.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig1_HTML.jpg
Fig. 1

In vivo bioluminescence images after inoculation of Ba/F3-Luc/Wt cells. The pseudocolour luminescent image (blue, green, yellow and red from the weakest to the strongest) is overlaid on the grey-scale photographic image. After subcutaneous inoculation in the right femoral region of the mouse, light emission is shown at the site of inoculation on the dorsal image (a). Following intravenous inoculation, the ventral (b) and left lateral (c) images reveal extensive light emission, particularly at the sites corresponding to the lung, liver and spleen

Luminescence in dilution series

To assess the relationship between bioluminescent signal intensity and viable cell numbers, we prepared a dilution series of the Ba/F3-Luc/Wt cells (range 2.5×104–1.6×106 cells/ml) and measured the luminescence for a given volume of cell suspension using the standard and intact-cell luciferase assays. For the standard luciferase assay, the luminescence increased in proportion to the increasing cell numbers (Fig. 2a) and the luminescence per cell was constant, irrespective of cell number (range 195.3–201.5 cps/cell). Luminescence measured by the intact-cell luciferase assay was also highly proportional to cell number (Fig. 2b), and the luminescence per cell was stable (range 1.467–1.571 cps/cell). All of the following assays were performed in the linear range.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig2_HTML.gif
Fig. 2

Relationship between cell number and luminescence in a cell dilution series. The levels of luminescence in the standard luciferase assay (a) and in the intact-cell luciferase assay (b) were highly proportional to the numbers of Ba/F3-Luc/Wt cells prepared by serial dilution. Error bars are not visible because the standard errors are too small

IL-3 levels and luciferase activities

We evaluated the effects of mIL-3 on cell proliferation and luciferase expression for factor-dependent Ba/F3-Luc cells, not expressing BCR-ABL. After 24-h incubation in the presence of different concentrations of mIL-3 (100, 10 and 1 pg/ml), viable cell numbers, luciferase activities in the standard luciferase assay and proliferation indices were determined. The viable cell number increased with increasing mIL-3 concentration, indicating dose-dependent stimulation of cell proliferation (Fig. 3a). Luminescence for a given volume of cell suspension also increased with increasing mIL-3 concentration (Fig. 3b). The dependence on mIL-3 concentration was more prominent for luminescence than for viable cell number, and thus luminescence per cell increased with increasing mIL-3 concentration (Fig. 3c), which suggests enhancement of luciferase expression by mIL-3. The proliferation index was higher for 100 pg/ml mIL-3, consistent with higher proliferative activity, than for 10 pg/ml or 1 pg/ml (Fig. 3d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig3_HTML.gif
Fig. 3

IL-3 concentration and luciferase activity. Factor-dependent Ba/F3-Luc cells (1×105 cells/ml) were seeded in a 24-well plate in the presence of different concentrations of mIL-3. After 24 h, the viable cell number (a), luminescence in the standard luciferase assay (b), luminescence per cell (c) and proliferation index (d) were determined for each well. IL-3 dependence was more pronounced for luminescence than for viable cell number, and luminescence per cell increased with increasing concentrations of mIL-3. Error bars in panels a, b and d represent standard errors

Monitoring of proliferation by luciferase assays

To investigate the validity of bioluminescent signal as a marker of cell proliferation, we evaluated the proliferation of factor-independent cells, Ba/F3-Luc/Wt cells and Ba/F3-Luc/Mut cells, by serial assessments of viable cell numbers, luminescence in the standard luciferase assay, luminescence in the intact-cell luciferase assay and proliferation indices. The culture medium was not changed after replating, and measurements were performed every 12 h. No substantial differences were found between the Ba/F3-Luc/Wt and Ba/F3-Luc/Mut cells (data not shown). The viable cell numbers increased rapidly and then reached a plateau (Fig. 4a). As for the standard luciferase assay, luminescence for a given volume of cell suspension increased during the proliferative phase, which suggests that this assay provides an indicator of cell proliferation. However, increases in luminescence tended to be less prominent than increases in viable cell number, implying a mild underestimation of proliferation by the standard luciferase assay. During the plateau phase, luminescence decreased despite constant viable cell numbers, and a discrepancy was noted between the proliferation assessed by viable cell counting and that assessed using the standard luciferase assay. Luminescence per cell for the standard luciferase assay showed a gradual reduction over time, suggesting an incubation time-dependent decline in luciferase expression (Fig. 4b). The cell cycle analysis also demonstrated a gradual decline in proliferation index, and the time course was similar between the luciferase activity per cell and the proliferative fraction (Fig. 4c). The increase in luminescence seen in the intact-cell luciferase assay was more pronounced than that in the standard luciferase assay and closely paralleled the increase in viable cell numbers during the proliferative phase (Fig. 4a). During the plateau phase, a discrepancy between viable cell counting and the intact-cell luciferase assay occurred to a lesser degree than between viable cell counting and the standard luciferase assay. The incubation time-dependent decrease in luminescence per cell for the intact-cell luciferase assay was evident but less prominent than that for the standard luciferase assay (Fig. 4b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig4_HTML.gif
Fig. 4

Monitoring of proliferation of Ba/F3-Luc/Wt cells by luciferase assays. The cells (5×104 cells/ml) were seeded in a 24-well plate, and measurements were performed every 12 h after a 12-h pre-incubation period. The viable cell numbers and luminescence values from the standard luciferase assay and from the intact-cell luciferase assay, expressed as percentages of baseline values, increased over incubation time (a), and the luciferase assays reflected cell proliferation during the proliferative phase. However, definite underestimation occurred during the plateau phase and was more pronounced for the standard luciferase assay than for the intact-cell luciferase assay. Luminescence per cell (b) and proliferation index (c) decreased over time. Error bars in panels a and c represent standard errors

Monitoring of responses to imatinib by luciferase assay

We assessed the effect of imatinib on cell proliferation and luciferase activity for factor-independent cells. After 24-h incubation of Ba/F3-Luc/Wt cells in the presence of 1.0, 0.5, 0.25 or 0 μM imatinib, viable cell counting, standard luciferase assay, and cell cycle analysis were performed. The presence of 1 μM imatinib mildly depressed the increase in viable cell numbers (Fig. 5a), while such an inhibitory effect was not evident with 0.5 μM or 0.25 μM imatinib. Luciferase activity assessed by the standard luciferase assay was reduced even at 0.25 μM when compared with the corresponding value in the absence of imatinib, and further decreased in a dose-dependent manner (Fig. 5b). The decrease in luciferase activity was more pronounced than that for viable cell number, and thus luminescence per cell was also reduced by the increasing imatinib concentration (Fig. 5c). The cell cycle analysis revealed dose-dependent decreases in the proliferative fraction (Fig. 5d). For the Ba/F3-Luc/Mut cells, which are resistant to imatinib, the presence of 1 μM imatinib had no substantial effect on viable cell number, luciferase activity, luminescence per cell or proliferative fraction (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig5_HTML.gif
Fig. 5

Imatinib concentration and luciferase activity. The Ba/F3-Luc/Wt cells (1×105 cells/ml) were seeded in a 24-well plate with different concentrations of imatinib. After 24 h, the viable cell number (a), luminescence in the standard luciferase assay (b), luminescence per cell (c) and proliferation index (d) were determined. Increases in imatinib concentration decreased all measures. The effect was more pronounced on luminescence than on viable cell number. Error bars in panels a, b and d represent standard errors

We sequentially assessed the proliferation of Ba/F3-Luc/Wt cells in the presence or absence of 1 μM imatinib. The measurements, including viable cell counting, standard luciferase assay and cell cycle analysis, were performed immediately after the addition of imatinib and every 12 h thereafter. Although time-dependent increases in viable cell number (Fig. 6a) and luminescence in the standard luciferase assay (Fig. 6b) were demonstrated in the presence and absence of imatinib, the increases were attenuated in the presence of imatinib. The inhibitory effect on viable cell number was not apparent 12 h after the addition of imatinib, but became evident at 24 h. The inhibitory effect on luciferase activity was evident even at 12 h and was more pronounced than that on cell number. Luminescence per cell showed a definite decline over time in the presence of imatinib, while it was almost constant in the absence of imatinib during the observation period (Fig. 6c). The proliferation index at 12 h was almost equal to that at baseline, irrespective of the presence or absence of imatinib (Fig. 6d). Otherwise, the time course for the proliferation index resembled that for luminescence per cell.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-005-0048-4/MediaObjects/259_2005_48_Fig6_HTML.gif
Fig. 6

Sequential assessments of Ba/F3-Luc/Wt cell proliferation in the presence or absence of 1 μM imatinib. Twelve hours after seeding the cells (2×104 cells) in 980-μl of medium in a 24-well plate, 20 μl of medium with or without imatinib was added. The measurements were performed immediately after the addition of the medium and every 12 h thereafter. The viable cell numbers (a) and luminescence in the standard luciferase assay (b) are expressed as percentages of the values at time 0. Imatinib inhibited the increase in cell number and, earlier and more severely, the increase in luciferase activity. Gradual reductions in luminescence per cell (c) and proliferative fraction (d) are apparent in the presence of imatinib. Error bars in panels a, b and d represent standard errors

Discussion

In vivo bioluminescence imaging offers a promising tool for small animal experiments. In this study, we generated p190-BCR-ABL-transduced Ba/F3 cells, which were either sensitive or resistant to imatinib, for monitoring by in vivo bioluminescence imaging. Since bioluminescent monitoring of tumour models requires stable expression of luciferase, we introduced the firefly luciferase gene under the control of the MMLV LTR, a representative constitutive promoter. No substantial differences in proliferation rate or responsiveness to imatinib were found between the cell lines with and without the luciferase gene, justifying the prediction of proliferation and treatment responses of cells that do not express luciferase, based on those of cells that stably express luciferase. Although it is possible to maintain stable gene expression in cell cultures using selection agents, stable expression of luciferase in the absence of selection pressure is needed for in vivo use. We confirmed the long-term stability of luciferase expression in medium not containing selection agents. We imaged mice using a CCD camera after subcutaneous or intravenous inoculation of the luciferase-expressing Ba/F3 cells and demonstrated the feasibility of visualising the cells, located either superficially or deeply, by in vivo bioluminescence imaging. In vivo light signal was clearly detected for wild-type mice coated with white fur as well as for nude mice. Luciferase expression in the cells is considered to be sufficient for in vivo imaging. These in vitro and in vivo results suggest that the cells established here have characteristics suitable for the bioluminescent evaluation of therapies in leukaemia model animal, whereas the feasibility of in vivo monitoring of disease progression and therapeutic response using the cells and bioluminescence imaging needs to be examined in future animal experiments.

Light emission from a tumour, as measured in bioluminescence imaging, is used as a quantitative marker of tumour burden [1, 2]. Proportionality between total luciferase activity of the tumour and the number of viable tumour cells, i.e. constancy of luciferase activity per viable cell, is desirable for such assessment. Previous cell culture studies demonstrated that viable cell numbers correlated linearly with light output in a dilution series of luciferase-expressing cells [37], and proportionality was confirmed in the present study for both the standard and intact-cell luciferase assays. However, for a dilution series, the cell conditions are uniform and the potential variations in luciferase expression related to changes in the physiological status of the cell are not taken into consideration. We cultured luciferase-expressing cells under various conditions and compared the luciferase activities, measured using the standard luciferase assay, with the viable cell numbers. In most cases, the time- and dose-dependent patterns for the luciferase activities were similar to those for the viable cell numbers, which suggests that changes in luciferase activity generally reflect changes in viable cell number. However, the luciferase activity per viable cell varied significantly according to the mIL-3 concentration for the factor-dependent cells, and according to the stage of proliferation and imatinib concentration for the factor-independent cells, which distorted the proportionality between viable cell number and bioluminescent signal intensity. Cytokines, proliferative stages and therapeutic drugs are suggested to affect luciferase expression, probably due to alterations in the activity of the LTR promoter. The activity of the CMV promoter has been reported to depend on the cell cycle stage and to be high during the S phase [20]. In our study employing the MMLV LTR, the luciferase activity per cell tended to be higher for cell cultures containing a larger fraction of proliferating cells. Although the molecular mechanisms remain to be studied, cytokines and cell culture conditions may have similar effects on proliferative activity and LTR activity. The signal on in vivo bioluminescence imaging may be related not only to viable cell numbers but also to proliferative activity.

Positive correlations between tumour burden and bioluminescent signal have been shown in many in vivo studies [3, 6, 819]. However, Scatena et al. described constant bioluminescent signals despite 3.3-fold increase in tumour volume [34]. A similar discrepancy between liver weight and bioluminescence has been demonstrated for hepatic tumour models with no evidence of significant necrosis or fibrosis on histological examinations [7]. While these observations are attributable partly to enhanced absorption of light photons within large tumours, changes in tumour physiology may also be responsible. In the monitoring of proliferation of the factor-independent, luciferase-expressing cells in the present study, the luciferase activity per viable cell decreased gradually, and the standard luciferase assay underestimated the proliferation, especially during the plateau phase. The medium was not changed during the observation period, and poor medium condition resulting from long incubation may have depressed the activity of the LTR promoter. The progression of implanted tumours in living animals can cause dynamic changes in the microenvironment, such as alterations in the blood supply and oxygenation. Such changes may influence the activity of the LTR promoter and, consequently, signal on in vivo bioluminescence imaging. In vivo bioluminescence imaging has been used for the evaluation of the effects of antineoplastic therapies, including imatinib treatment [28], and bioluminescence has also been used for the in vitro assessment of therapeutic response [13, 35, 36]. In studies with imatinib-sensitive cells, the addition of imatinib to the culture medium reduced the luciferase activity per viable cell in a dose-dependent manner. In sequential measurements after exposure to imatinib, inhibition of the increase in luciferase activity occurred earlier and was more severe than that seen in the viable cell numbers, which suggests that luciferase expression declines early after the addition of imatinib owing to a reduction in the activity of the LTR promoter. Residual tumour burden may be underestimated when the therapeutic effect of imatinib is assessed by in vivo bioluminescence imaging. From another viewpoint, the reduction in luciferase activity per cell may be beneficial, since it enables more sensitive detection of the therapeutic effect.

We performed serial assessments of proliferation using the intact-cell luciferase assay as well as the standard luciferase assay. Although the standard luciferase assay appears to accurately evaluate luciferase activity, the signal on in vivo bioluminescence imaging is not dependent solely on luciferase activity. For in vivo imaging, D-luciferin is absorbed through the peritoneum, is delivered by the blood flow and enters the luciferase-expressing cells. D-Luciferin is oxidised by luciferase in the presence of co-factors (oxygen, adenosine triphosphate and magnesium), resulting in light emission. Some of the emitted light photons pass through the tissues and, finally, are detected by the CCD camera. The intensity of the signal measured by in vivo imaging may depend on various factors, such as D-luciferin absorption through the peritoneum, blood flow, cell membrane permeability, the availability of co-factors, intracellular pH and the transparency of overlying tissues, in addition to the amount of luciferase. Among these parameters, the importance of attenuation of emitted light by overlying tissues has been well recognised [4]. It has also been pointed out that cell uptake of D-luciferin is inefficient and may be a limiting factor for in vivo bioluminescence [37]. The intact-cell luciferase assay may be affected by cell membrane permeability and the intracellular environment and appears to simulate in vivo imaging more faithfully than the standard luciferase assay. In the monitoring of cell proliferation, the decrease in luminescence per cell was less prominent for the intact-cell luciferase assay than for the standard assay. Prolonged incubation appears to lower the pH of the culture medium, thereby reducing the negative charge of D-luciferin added to the medium. The relative preservation of luminescence per cell in the intact-cell luciferase assay may be attributable to enhancement of cell membrane permeability to D-luciferin by the low pH of the medium [38, 39]. Variations in D-luciferin availability due to changes in the tissue microenvironment may influence the intensity of the signal on in vivo bioluminescence imaging. Various changes may occur in association with disease progression and therapeutic responses in living mice and may affect the relationship between bioluminescent signal and tumour burden. The relationship in living mice remains to be investigated under various conditions.

In conclusion, we generated p190 BCR-ABL-transformed Ba/F3 cell lines stably expressing luciferase under the control of a retroviral LTR for in vivo evaluations of treatment strategies for leukaemia. Our cell culture studies indicate that the bioluminescent signal generally reflects cell proliferation and responses to imatinib. However, differences in cell culture conditions and the addition of imatinib alter the levels of luminescence per cell as well as the fraction of proliferating cells. Although in vivo bioluminescence imaging would allow non-invasive monitoring of leukaemia model animals, environmental factors and therapeutic interventions may cause discrepancies between tumour burden and the intensity of bioluminescence in relation to changes in proliferative activity. The association of luciferase expression with proliferative activity may enhance the sensitivity of bioluminescence imaging to therapeutic responses. The relationship between viable cell number and bioluminescence may vary depending on the cell types and promoters, and it is recommended to examine the relationship for each luciferase-expressing cell line.

Acknowledgements

This work was supported, in part, by grants-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Copyright information

© Springer-Verlag 2006